The embers of collective intelligence in science

In this blog post, I’ll explore reasons to call science a product of collective intelligence. I’ll look for something more than the sum of its parts, something more than a Cartesian or Baconian view of science as an edifice built by the contributions of many individuals. Are there steps of scientific progress taken by several minds together? Truly irreducible achievements of collective intelligence?

Collective intelligence may evoke associations with the wisdom of the crowd. This often refers to social processes that converge to the majority opinion/decision and thereby manage to surpass the population average, or the crowd average gets closer to a true value than the standard deviation. Since science mostly doesn’t proceed by choosing between alternatives or estimating numbers, it feels like a far-fetched effect to look for. Well, maybe there is some wisdom of the crowd in funding organizations, but probably more herd mentality and groupthink. (I also remember hearing about some organization voting for changing the exponent of Kleiber’s law, but I can’t find any evidence of that.)

Wisdom of the crowd aside, the first thing that comes to my mind is the collective process of conceptualization—how we together decide which phenomena deserve a term of their own. The most remarkable example here, I think, is thermodynamics. Think of the potential functions—enthalpy, the various free energies, the chemical potential—all of which are abstract concepts that we can’t directly measure in the lab, let alone sense. These were put together into an unusually self-consistent and self-contained theory by a bunch of steam engineers who probably didn’t do much more than play with their steam engines.

But then again, there are too many examples where this process simply never worked, especially when it comes to epistemological, meta-scientific conceptualization. Think of the ambiguities of what “theory,” “model,” “evidence,” “law,” “mechanism,” etc., mean. These words vary not only between the disciplines but sometimes even at a smaller granularity.

Another process that just doesn’t happen within a single brain is when we collectively decide a concept is useful, but then the concept is so abstract and nebulous that people continue to develop their own motivations, definitions, and concretizations of it. This kind of collective reification could be good. Think of all the influential theories that sprang out of people trying to wrap their minds around “entropy.” (Yes, thermodynamics again!) Before Boltzmann, “entropy” was almost tautological—”that something which never decreases in a thermodynamic process.” Then Gibbs hit upon a mathematical form that would go on to be foundational for information theory, compression, cryptography, error detection, etc. But he might never have done that without the peer pressure of considering entropy central. Entropy was like a kōan for the scientific hivemind.

But then again, we often fall in love with deceptive concepts, sending us down cul-de-sacs: phlogistons, string theory, the optimization principles of non-equilibrium thermodynamics, self-organized criticality, machine consciousness, you name it. In general, it’s hard to judge the collective processes setting the priorities among scientific topics, but I have a feeling they hinder more than they spur the advancement of science.

What other manifestations of collective intelligence are there in science? At a microscopic level, there is, of course, the effervescence of teamwork when we can leverage a diverse set of expertise to write papers no one alone could write. However, this is not a population-wide emergent property; nothing pointing to us being neurons in a brain called science or the constituents of a sphere of pure reason like the living make up the biosphere (as once philosophized in the grandiloquent writings of the French Jesuit priest Pierre Teilhard de Chardin and the Russian polymath Vladimir Vernadsky).

To summarize, I see glimmers of collective intelligence but ever so much more collective stupidity and wholes that are the sum of their parts. Is there a way to conjure those magic moments?

Shout out to Marko Jusup, Sune Lehmann, and Taha Heidari whose comments and chats gave several ideas to this post. The cover picture is a color-inverted Smith chart from here.